EnrichGT Documents

Author

Zhiming Ye

Published

December 18, 2024

Overview

The functions from EnrichGT are starting with “egt_” or “database_”.

graph LR
    subgraph Enrichment Analysis
        A[egt_enrichment_analysis]
        B[egt_gsea_analysis]
    end

    subgraph Pathway Databases
        D[database_* funcs]
    end

    subgraph Visualize results
        P1[egt_plot_results]
        P2[egt_plot_umap]
    end

    subgraph egt_recluster_analysis
        K1[Pretty table]
        CC[cluster modules]
        MG[gene modules]
    end

    subgraph Pathway Act. and TF infer 
        
        I[egt_infer]
    end

    D --> A
    D --> B

    A --> C[Enriched Result]
    B --> C

    C --> CC
    C --> MG

    C --> P1

    CC --> K1
    MG --> K1

    CC --> P1
    CC --> P2

    MG --> I

Install EnrichGT

install.packages("pak")
pak::pkg_install("ZhimingYe/EnrichGT")

install.packages("devtools")
library(devtools)
install_github("ZhimingYe/EnrichGT")

The AnnotationDbi, fgsea, reactome.db and GO.db were from BioConductor and might be slow to install. If you can’t install, please re-check your web connections or update your R and BioConductor, or use Posit Package Manager to install when using old R.

Meet EnrichGT

Important

See package function page for further information (For example, how to use a function)

You can use ? function to get further help when installed.

?egt_recluster_analysis

Core Function

Enrichment of genes

This is a C++ accelerated over representation analysis tool. The only things you need is your favourite gene symbols. If is all prepared, then load a database, run it!

The difference from other tools

Compared to the most popular clusterProfiler, the functions of EnrichGT differ slightly. This is mainly to accommodate wet lab researchers. First, most beginners are confused by the default input of clusterProfiler, which is “ENTREZ ID.” Most people familiar with biology are used to Gene Symbols, and even Ensembl IDs are not widely known, let alone a series of seemingly random numbers. Therefore, EnrichGT uses Gene Symbol as the default input, seamlessly integrating with most downstream results from various companies, making it more suitable for non-experts in the lab.

Second, clusterProfiler outputs an S4 object, which may be too complex for beginners (this is no joke); whereas EnrichGT outputs a simple table. The time of non-experts is precious, so I made these two adjustments. The only downside is that the GSEA peak plot is difficult to generate, but in reality, we focus more on NES and p-values, and in this case, bar plots are more convincing.

Furthermore, The pre-processing step of the hypergeometric test in EnrichGT’s ORA function (which determines overlap) is accelerated using hash tables in C++, making it over five times faster than clusterProfiler::enricher(), which is a pure R implementation.

res <- egt_enrichment_analysis(genes = DEGtable$Genes,
database = database_GO_BP())

res <- egt_enrichment_analysis(genes = c("TP53","CD169","CD68","CD163",
                                         "You can add more genes"),
database = database_GO_ALL())

res <- egt_enrichment_analysis(genes = c("TP53","CD169","CD68","CD163",
                                         "You can add more genes"),
database = database_from_gmt("MsigDB_Hallmark.gmt"))

Example of ORA

library(dplyr)
library(tibble)
library(org.Hs.eg.db)
library(gt)
library(testthat)
library(withr)
library(EnrichGT)
library(readr)

DEGexample <- read_csv("./DEG.csv")

New names:
Rows: 15903 Columns: 7
── Column specification
──────────────────────────────────────────────────────── Delimiter: "," chr
(1): ...1 dbl (6): baseMean, log2FoldChange, lfcSE, stat, pvalue, padj
ℹ Use `spec()` to retrieve the full column specification for this data. ℹ
Specify the column types or set `show_col_types = FALSE` to quiet this message.
• `` -> `...1`

DEGexample_UpReg <- DEGexample |> dplyr::filter(pvalue<0.05,log2FoldChange>0.7)
ora_result <- egt_enrichment_analysis(genes = DEGexample_UpReg$...1,database = database_GO_BP(org.Hs.eg.db))


✔ success loaded database, time used : 17.0558180809021

head(ora_result)

          ID                                        Description GeneRatio
1 GO:0035249               synaptic transmission, glutamatergic    19/457
2 GO:0051966 regulation of synaptic transmission, glutamatergic    16/457
3 GO:0050804       modulation of chemical synaptic transmission    39/457
4 GO:0099177             regulation of trans-synaptic signaling    39/457
5 GO:0050808                               synapse organization    36/457
6 GO:0048168         regulation of neuronal synaptic plasticity    12/457
    BgRatio       pvalue     p.adjust       qvalue
1 111/18870 2.055491e-11 8.053235e-08 8.409591e-05
2  79/18870 5.742011e-11 8.053235e-08 8.409591e-05
3 489/18870 7.257625e-11 8.053235e-08 8.409591e-05
4 490/18870 7.711980e-11 8.053235e-08 8.409591e-05
5 483/18870 2.512088e-09 2.098598e-06 1.109159e-03
6  56/18870 7.497368e-09 4.183196e-06 1.109159e-03
                                                                                                                                                                                                                                  geneID
1                                                                                                                ATP1A2/GRIA4/GRID2/GRIK2/GRIK3/GRIN1/GRIN2A/GRIN2B/GRIN2D/GRM1/GRM5/GRM8/DGKI/NRXN1/NLGN1/UNC13A/MAPK8IP2/CACNG5/UNC13C
2                                                                                                                                   ATP1A2/GRIK2/GRIK3/GRIN1/GRIN2A/GRIN2B/GRIN2D/GRM1/GRM5/GRM8/DGKI/NRXN1/NLGN1/UNC13A/MAPK8IP2/CACNG5
3 ACHE/APOE/ATP1A2/CA2/CAMK2B/CDC20/GFAP/GRIA4/GRID2/GRIK2/GRIK3/GRIN1/GRIN2A/GRIN2B/GRIN2D/GRM1/GRM5/GRM8/HRAS/MAP1B/NTRK2/SLC6A1/CNTN2/VGF/WNT5A/INA/DGKI/DLGAP1/NRXN1/RIMS3/NMU/NLGN1/UNC13A/MAPK8IP2/ERC2/CACNG5/LRFN2/UNC13C/SHISA9
4 ACHE/APOE/ATP1A2/CA2/CAMK2B/CDC20/GFAP/GRIA4/GRID2/GRIK2/GRIK3/GRIN1/GRIN2A/GRIN2B/GRIN2D/GRM1/GRM5/GRM8/HRAS/MAP1B/NTRK2/SLC6A1/CNTN2/VGF/WNT5A/INA/DGKI/DLGAP1/NRXN1/RIMS3/NMU/NLGN1/UNC13A/MAPK8IP2/ERC2/CACNG5/LRFN2/UNC13C/SHISA9
5             ACHE/APOE/KIF1A/CAMK2B/CDC20/CDH6/CTNNA2/DSCAM/GAP43/GRID2/GRIN2B/GRM5/MAP1B/NRCAM/NTRK2/RAC3/SIX1/SLC6A1/CNTN2/WNT5A/INA/NRXN1/NLGN1/UNC13A/ERC2/IL1RAPL2/SEZ6L2/TREM2/LRFN2/IGSF9/BCAN/SYNDIG1/DNER/ADGRF1/LHFPL4/UNC13C
6                                                                                                                                                                APOE/CAMK2B/GRIK2/GRIN1/GRIN2A/GRIN2B/GRIN2D/GRM5/HRAS/CNTN2/VGF/SHISA9
  Count
1    19
2    16
3    39
4    39
5    36
6    12

Have many sources of genes?

This function also support many groups of genes, you can input a list.

# For many groups of genes
res <- egt_enrichment_analysis(list(Macrophages=c("CD169","CD68","CD163"),
Fibroblast=c("COL1A2","COL1A3"),"You can add more groups"),
 database = database_from_gmt("panglaoDB.gmt"))

Enrichment of weighted genes (GSEA)

Genes with specific weights (e.g. the log2FC) can use GSEA method. It should input a pre-ranked geneset. This use C++ accelerated fgsea::fgsea() as backend, so it is also very fast.

How to build pre-ranked gene set?

genes_with_weights(genes,weights) function is used to build the pre-ranked gene set for GSEA analysis.

# From DEG analysis Results
res <- egt_gsea_analysis(genes = 
                           genes_with_weights(genes = DEG$genes, 
                                              weights = DEG$log2FoldChange),
                         database = database_GO_BP()
                         )

# From PCA
res <- egt_gsea_analysis(genes = genes_with_weights(genes = PCA_res$genes,
                                                    weights =PCA_res$PC1_loading),
                         database = database_from_gmt("MsigDB_Hallmark.gmt")
                         )

Featured Function

Enrichment of Enriched Results

The enriched result is too messy? Clean up it!

Form clusterProfiler ?

This can also supports the results from clusterProfiler, so you can use any tool to do this.

Why the re-enrichment is necessary?

Challenges in Biological Gene Enrichment Analysis

Gene enrichment analysis can often be misleading due to the redundancy within gene set databases and the limitations of most enrichment tools. Many tools, by default, only display a few top results and fail to filter out redundancy. This can result in both biological misinterpretation and valuable information being overlooked.

For instance, high expression of certain immune genes can cause many immune-related gene sets to appear overrepresented. However, a closer look often reveals that these gene sets are derived from the same group of genes, which might represent only a small fraction (less than 10%) of the differentially expressed genes (DEGs). What about the other 90%? Do they hold no biological significance?

Current Solutions

clusterProfiler is one of the most powerful tools in R for enrichment analysis. It’s designed with pathway redundancy in mind and includes the clusterProfiler::simplify function to address this issue. This method, based on GOSemSim for GO similarity evaluation, is scientifically robust and highly effective.

However, there are some drawbacks:

GOSemSim is not fast, particularly when dealing with large or complex gene sets.
It doesn’t support databases like KEGG or Reactome.

Using GOSemSim to measure the semantic similarity between pathways is, theoretically, the best way to tackle redundancy. However, in practical cases—especially in experimental bioinformatics validation—researchers are more focused on the genes behind these pathways rather than the pathways themselves.

Alternative Approaches

Although clustering pathways based on gene overlap has received some criticism, it remains a viable approach in many situations. For this reason, I developed BioThemeFinder a few years ago to solve this problem. However, the tool is so awful (I am poor in coding…)

Today, two excellent alternatives exist:

simplifyEnrichment: This package is more scientifically rigorous (based on semantic similarity) and creates beautiful visualizations. It also doesn’t support databases like KEGG or Reactome.
aPEAR: A simpler and faster tool that better aligns with practical needs, making it my preferred choice.

However, both of these tools have a common limitation: their visualizations are optimized for publication purposes rather than for exploratory research. I often find myself exporting CSV files or struggling with RStudio’s preview pane to fully explore enrichment tables. This inspired me to develop a more efficient solution. Also, they are slow.

Goals of This Package

The main purpose of developing this package is to provide a lightweight and practical solution to the problems mentioned above. Specifically, this package aims to:

Cluster enrichment results based on hit genes or core enrichment from GSEA using term frequency analysis (from the output of the powerful clusterProfiler). This provides a clearer view of biological relevance by focusing on the genes that matter most.

# From results generated before
res <- egt_enrichment_analysis(genes = DEGtable$Genes,
database = database_GO_BP())

re_enrichment_results <- egt_recluster_analysis(
  res,
  ClusterNum = 17,
  P.adj = 0.05,
  force = F,
  nTop = 10,
  method = "ward.D2"
)

You can see the structure of egt_obj. The first slot is the result table, and the second slot contains gt table.

str(re_enrich,max.level = 2)

Formal class 'EnrichGT_obj' [package "EnrichGT"] with 7 slots
  ..@ enriched_result     : tibble [103 × 7] (S3: tbl_df/tbl/data.frame)
  ..@ gt_object           :List of 17
  .. ..- attr(*, "class")= chr [1:2] "gt_tbl" "list"
  ..@ gene_modules        :List of 16
  ..@ pathway_clusters    :List of 16
  ..@ document_term_matrix:Formal class 'dgCMatrix' [package "Matrix"] with 6 slots
  ..@ clustering_tree     :List of 7
  .. ..- attr(*, "class")= chr "hclust"
  ..@ raw_enriched_result :'data.frame':    175 obs. of  7 variables:

HTML reports (gt table)

Also, because of the messy result table is hardly to read, EnrichGT help you convert it into pretty gt HTML tables. This only supports the re-enriched results.

The gt_object is a pure object of gt package, you can use any function on it, like:

re_enrichment_results@gt_object |> gt_save("test.html") # Save it use basic gt functions.

For further usage of gt package, please refer to https://gt.rstudio.com/articles/gt.html.

See re-enrichment example for further demo.

Ploting functions

Warning

The Dot Plot supports simple enrichment result data.frame and re-enriched egt_object, but UMAP plot only supports re-enriched egt_object.

HTML gt table satisfied most of things, but for others. Though we don’t want this package become complex (i.e., you can simple draw your figure using ggplot2 for enriched tables by yourself.) But we still provide limited figure ploting functions.

Dot Plot

egt_plot_results(re_enrich)

UMAP Plot

egt_plot_umap(re_enrich)

Warning: ggrepel: 18 unlabeled data points (too many overlaps). Consider
increasing max.overlaps

DataBases Helpers

How to specify species?

EnrichGT use AnnotationDbi for this. you can use org.Hs.eg.db for human and org.Mm.eg.db for mouse. For others, please refer to BioConductor.

But for non-AnnotationDbi source database, you do not need to provide this, like database_CollecTRI_human() return database about human only.

Built in database form AnnotationDbi

You should add argument OrgDB for fetching them.

Example:

database_GO_BP(OrgDB = org.Hs.eg.db)

GO Database

database_GO_BP(), database_GO_CC(), database_GO_MF(), database_GO_ALL()

Reactome Database

database_Reactome()

Progeny Database

For pathway activity infer, database_progeny_human() and database_progeny_mouse()

CollecTRI Database

For Transcript Factors infer, database_CollecTRI_human() and database_CollecTRI_mouse()

Read Addition Gene Sets from local

EnrichGT supports reading GMT files, You can obtain GMT files from MsigDB.

database_from_gmt("Path_to_your_Gmt_file.gmt")

Where is KEGG?

KEGG limited the commercial usage. And you should use the KEGG REST API to download it. I have no time to achieve it now. But you can use KEGG Database from MsigDB instead (KEGG_MED and KEGG_Classical).

Reading is slow?

From 0.5.0, EnrichGT implemented a cache system. So when load a same database the second time, it will be much faster.

test <- database_GO_MF(org.Hs.eg.db)

✔ success loaded database, time used : 6.91346096992493

test_reload <- database_GO_MF(org.Hs.eg.db)

✔ Use cached database: GO_MF_org.Hs.eg.db

Future development plan

Version 0.5 is freeze in Dec 19th 2024.

Version 0.6 targets (Will be starting in 2024-12-25)

support KEGG online reading
Add simple network plot like cnetplot
Add drawing function for egt_infer()
Better dot plot
Self-built gene converter

Acknowledgement

This package is inspired by famous clusterProfiler. But since 0.5 version, the major enrichment functions of EnrichGT have replaced by the self-implemented functions, which provides a much light-weight experience. But without clusterProfiler, I won’t try to write this package.

If also using clusterProfiler ?

Please cite:

T Wu^#, E Hu^#, S Xu, M Chen, P Guo, Z Dai, T Feng, L Zhou, W Tang, L Zhan, X Fu, S Liu, X Bo^*, G Yu^*. clusterProfiler 4.0: A universal enrichment tool for interpreting omics data. The Innovation. 2021, 2(3):100141. doi: 10.1016/j.xinn.2021.100141